ISU and Ares use AI to develop polymers for solid-state batteries
ISU Specialty Chemical and Ares Materials are jointly developing polymer materials for solid-state batteries. The collaboration also aims to create artificial intelligence tools that accelerate formula optimization using battery cell-level test data.
- ISU and Ares are developing polymers for solid-state batteries.
- Ares's AI framework selects the next experiments using cell-level performance data.
- ISU is responsible for manufacturing and commercial supply.
- The development aims to shorten the formula optimization cycle.
- Models generated from development of the first material will also be used for subsequent generations.
Materials tuned using cell test data
ISU Specialty Chemical, the lithium sulfide-producing subsidiary of South Korea's ISU Group, and Ares Materials, which is based in Plano, United States, and Guadalajara, Mexico, announced their collaboration on November 3, 2026. The companies are developing polymer materials for use in solid-state batteries, along with an artificial intelligence framework for optimizing them.
Under the agreement, Ares will develop the polymer formulations and the AI framework that tunes them based on cell-level battery test data. ISU will contribute its expertise in specialty chemicals and lithium sulfide production, and will also lead manufacturing and commercial supply.
The goal is to shorten development cycles
Developing polymers for solid-state cells typically requires new test cells to be built and cycled for every formulation change. Battery manufacturers say this creates lengthy development cycles, while making it difficult to determine precisely which component or manufacturing step contributed to a result.
The situation is further complicated by changes to cell specifications as the technology evolves. A formulation already qualified for a particular target must therefore be re-evaluated when requirements change.
AI tools for materials generally fall into two broad categories. Discovery systems search for new molecules, while solutions used for formula optimization adjust the composition and processing of a selected chemistry to meet the targets required in a working device. The ISU and Ares collaboration focuses on the latter area.
According to Ares, the framework learns from measured, device-level performance data. It breaks down individual test results according to the contribution of composition and process variables, then uses this information to select the next experiments. The companies said that the performance data and models generated during development of the first material will also be usable for subsequent product generations. This means new developments can start from an existing performance model.
ISU to handle manufacturing and supply
ISU Specialty Chemical's revenue in 2025 was approximately 411.5 billion South Korean won, equivalent to around $290 million, according to the announcement. The company manufactures molecular-weight regulators, including TDM, NOM and NDM, and completed its commercial lithium sulfide plant this year.
Founded in 1969, ISU Group is headquartered in Seoul. Its three publicly listed subsidiaries generated combined revenue of more than 3 trillion South Korean won in 2025, equivalent to around $2.2 billion, according to the announcement.
“Lithium sulfide gives us a position in the solid-state battery supply chain. This partnership is about what connects to it,” said Lyu Seung-Ho, CEO of ISU Specialty Chemical. He added that each generation developed through the partnership will make the next one faster, and that customers in South Korea, Japan and the United States are asking for this capability.
Partnership promises faster development for users
Ares Materials is developing polymers and a closed-loop AI infrastructure for formulating them against real-world performance. The company was selected for the Apple Impact Accelerator and Amazon Climate Tech Accelerator programs.
“Every materials supplier knows this pattern: we reach the specification, the specification changes, and we start over,” said Dr. David Arreaga, CEO of Ares Materials. He said the system being developed with ISU aims to break this cycle. If the model already knows how individual formulation variables affect device performance, meeting a new target can involve a modification rather than a complete restart.
Liz Lee, associate director at Counterpoint Research, said that combining AI-based materials intelligence with established specialty-chemical manufacturing could help better materials be developed and brought to market with fewer iterations and in less time. According to the announcement, quality, cost and scalability are all important considerations for next-generation battery materials.
